Training in progress, epoch 2, checkpoint
Browse files- last-checkpoint/global_step626/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +1 -1
- last-checkpoint/global_step626/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt +1 -1
- last-checkpoint/global_step626/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt +1 -1
- last-checkpoint/global_step626/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt +1 -1
- last-checkpoint/latest +1 -1
- last-checkpoint/model-00001-of-00004.safetensors +1 -1
- last-checkpoint/model-00002-of-00004.safetensors +1 -1
- last-checkpoint/model-00003-of-00004.safetensors +1 -1
- last-checkpoint/model-00004-of-00004.safetensors +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +483 -2
last-checkpoint/global_step626/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16060527788
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a2cbc934df001ac2a9ccfa1c3d659c01006a6b033fcf74c5531c89718e0ff73a
|
| 3 |
size 16060527788
|
last-checkpoint/global_step626/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16060527788
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5a5689f78e8a6d51e413284fde8f84af46c4e0061e29126b1d87c79d272edb86
|
| 3 |
size 16060527788
|
last-checkpoint/global_step626/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16060527788
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed1fb7d533c66f7293b771422adab0ae419f7aa890395d0ecd127cab35e4c6f9
|
| 3 |
size 16060527788
|
last-checkpoint/global_step626/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16060527788
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f65e6f9c60b31e8caa69dcc7c8b512638a97c3ab5179bb9d0b47099666b8ac6d
|
| 3 |
size 16060527788
|
last-checkpoint/latest
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
|
|
|
|
| 1 |
+
global_step626
|
last-checkpoint/model-00001-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4976698672
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2dc4dedb86acccf2e1c5b4a75ed6634cc7d2f756d42627de228881233991669
|
| 3 |
size 4976698672
|
last-checkpoint/model-00002-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999802720
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5462a38b49bba85e5ad8b66745f7c1a636c22c4ba050213c99d26a28b002a6b5
|
| 3 |
size 4999802720
|
last-checkpoint/model-00003-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4915916176
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:31e156640c29ba1aad1d24dec864f5b9b9e3a691befaaff77044f6811549a480
|
| 3 |
size 4915916176
|
last-checkpoint/model-00004-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1168138808
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a1dd80c6bba1a45d169fc29d48de0878da37846e60bf945be15eaee3dc9c30bd
|
| 3 |
size 1168138808
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2307c03867cef25b5028feb9a23f80e784b9af9a615de13ddca560a6a90fb593
|
| 3 |
size 15024
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:50218cfaecdd818354e567b7167c13899e3b42297e7d8f58bd7e732cfa547800
|
| 3 |
size 15024
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9814a66b49861f5495b06dae3be12ddf7185b88e2cae1fb808ca9efd99d5807f
|
| 3 |
size 15024
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a7be93040a65e0a29975f6c70b94418e1fdf88423a50c58aa572141d3c92fbfc
|
| 3 |
size 15024
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7725b44c69725321786f1e58dccd7ea4d3ae5794ea47bd49c0d4a139aec266d
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -503,6 +503,487 @@
|
|
| 503 |
"eval_samples_per_second": 15.872,
|
| 504 |
"eval_steps_per_second": 0.556,
|
| 505 |
"step": 313
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 506 |
}
|
| 507 |
],
|
| 508 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 626,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 503 |
"eval_samples_per_second": 15.872,
|
| 504 |
"eval_steps_per_second": 0.556,
|
| 505 |
"step": 313
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.0223642172523961,
|
| 509 |
+
"grad_norm": 27.543304377009523,
|
| 510 |
+
"learning_rate": 3.662721893491124e-07,
|
| 511 |
+
"logits/chosen": -1.65625,
|
| 512 |
+
"logits/rejected": -1.703125,
|
| 513 |
+
"logps/chosen": -240.0,
|
| 514 |
+
"logps/rejected": -258.0,
|
| 515 |
+
"loss": 0.4082,
|
| 516 |
+
"rewards/accuracies": 0.875,
|
| 517 |
+
"rewards/chosen": -0.03759765625,
|
| 518 |
+
"rewards/margins": 1.0546875,
|
| 519 |
+
"rewards/rejected": -1.09375,
|
| 520 |
+
"step": 320
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.0543130990415335,
|
| 524 |
+
"grad_norm": 27.327705966461778,
|
| 525 |
+
"learning_rate": 3.603550295857988e-07,
|
| 526 |
+
"logits/chosen": -1.6484375,
|
| 527 |
+
"logits/rejected": -1.65625,
|
| 528 |
+
"logps/chosen": -242.0,
|
| 529 |
+
"logps/rejected": -262.0,
|
| 530 |
+
"loss": 0.3683,
|
| 531 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 532 |
+
"rewards/chosen": -0.345703125,
|
| 533 |
+
"rewards/margins": 1.125,
|
| 534 |
+
"rewards/rejected": -1.46875,
|
| 535 |
+
"step": 330
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.0862619808306708,
|
| 539 |
+
"grad_norm": 17.914371176661845,
|
| 540 |
+
"learning_rate": 3.544378698224852e-07,
|
| 541 |
+
"logits/chosen": -1.6484375,
|
| 542 |
+
"logits/rejected": -1.703125,
|
| 543 |
+
"logps/chosen": -245.0,
|
| 544 |
+
"logps/rejected": -256.0,
|
| 545 |
+
"loss": 0.3574,
|
| 546 |
+
"rewards/accuracies": 0.887499988079071,
|
| 547 |
+
"rewards/chosen": -0.421875,
|
| 548 |
+
"rewards/margins": 1.1484375,
|
| 549 |
+
"rewards/rejected": -1.5703125,
|
| 550 |
+
"step": 340
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 1.1182108626198084,
|
| 554 |
+
"grad_norm": 27.18764754676586,
|
| 555 |
+
"learning_rate": 3.485207100591716e-07,
|
| 556 |
+
"logits/chosen": -1.703125,
|
| 557 |
+
"logits/rejected": -1.7109375,
|
| 558 |
+
"logps/chosen": -246.0,
|
| 559 |
+
"logps/rejected": -262.0,
|
| 560 |
+
"loss": 0.3483,
|
| 561 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 562 |
+
"rewards/chosen": -0.3046875,
|
| 563 |
+
"rewards/margins": 1.390625,
|
| 564 |
+
"rewards/rejected": -1.6953125,
|
| 565 |
+
"step": 350
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 1.1501597444089458,
|
| 569 |
+
"grad_norm": 16.652754238092474,
|
| 570 |
+
"learning_rate": 3.42603550295858e-07,
|
| 571 |
+
"logits/chosen": -1.65625,
|
| 572 |
+
"logits/rejected": -1.671875,
|
| 573 |
+
"logps/chosen": -245.0,
|
| 574 |
+
"logps/rejected": -260.0,
|
| 575 |
+
"loss": 0.3208,
|
| 576 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 577 |
+
"rewards/chosen": -0.18359375,
|
| 578 |
+
"rewards/margins": 1.4453125,
|
| 579 |
+
"rewards/rejected": -1.625,
|
| 580 |
+
"step": 360
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 1.182108626198083,
|
| 584 |
+
"grad_norm": 21.849555306547938,
|
| 585 |
+
"learning_rate": 3.366863905325444e-07,
|
| 586 |
+
"logits/chosen": -1.6875,
|
| 587 |
+
"logits/rejected": -1.6875,
|
| 588 |
+
"logps/chosen": -237.0,
|
| 589 |
+
"logps/rejected": -248.0,
|
| 590 |
+
"loss": 0.3809,
|
| 591 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 592 |
+
"rewards/chosen": -0.228515625,
|
| 593 |
+
"rewards/margins": 0.9296875,
|
| 594 |
+
"rewards/rejected": -1.15625,
|
| 595 |
+
"step": 370
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.2140575079872205,
|
| 599 |
+
"grad_norm": 22.779743163941504,
|
| 600 |
+
"learning_rate": 3.3076923076923075e-07,
|
| 601 |
+
"logits/chosen": -1.6796875,
|
| 602 |
+
"logits/rejected": -1.71875,
|
| 603 |
+
"logps/chosen": -246.0,
|
| 604 |
+
"logps/rejected": -258.0,
|
| 605 |
+
"loss": 0.3479,
|
| 606 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 607 |
+
"rewards/chosen": -0.177734375,
|
| 608 |
+
"rewards/margins": 1.1171875,
|
| 609 |
+
"rewards/rejected": -1.296875,
|
| 610 |
+
"step": 380
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 1.2460063897763578,
|
| 614 |
+
"grad_norm": 17.054365580798475,
|
| 615 |
+
"learning_rate": 3.2485207100591716e-07,
|
| 616 |
+
"logits/chosen": -1.65625,
|
| 617 |
+
"logits/rejected": -1.7109375,
|
| 618 |
+
"logps/chosen": -237.0,
|
| 619 |
+
"logps/rejected": -252.0,
|
| 620 |
+
"loss": 0.3299,
|
| 621 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 622 |
+
"rewards/chosen": 0.0015106201171875,
|
| 623 |
+
"rewards/margins": 1.28125,
|
| 624 |
+
"rewards/rejected": -1.28125,
|
| 625 |
+
"step": 390
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.2779552715654952,
|
| 629 |
+
"grad_norm": 20.840224552737123,
|
| 630 |
+
"learning_rate": 3.189349112426035e-07,
|
| 631 |
+
"logits/chosen": -1.6953125,
|
| 632 |
+
"logits/rejected": -1.59375,
|
| 633 |
+
"logps/chosen": -241.0,
|
| 634 |
+
"logps/rejected": -264.0,
|
| 635 |
+
"loss": 0.3656,
|
| 636 |
+
"rewards/accuracies": 0.875,
|
| 637 |
+
"rewards/chosen": -0.2265625,
|
| 638 |
+
"rewards/margins": 1.15625,
|
| 639 |
+
"rewards/rejected": -1.3828125,
|
| 640 |
+
"step": 400
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.3099041533546325,
|
| 644 |
+
"grad_norm": 24.15916736976353,
|
| 645 |
+
"learning_rate": 3.130177514792899e-07,
|
| 646 |
+
"logits/chosen": -1.6328125,
|
| 647 |
+
"logits/rejected": -1.6796875,
|
| 648 |
+
"logps/chosen": -243.0,
|
| 649 |
+
"logps/rejected": -258.0,
|
| 650 |
+
"loss": 0.3516,
|
| 651 |
+
"rewards/accuracies": 0.824999988079071,
|
| 652 |
+
"rewards/chosen": -0.00360107421875,
|
| 653 |
+
"rewards/margins": 1.1328125,
|
| 654 |
+
"rewards/rejected": -1.140625,
|
| 655 |
+
"step": 410
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.34185303514377,
|
| 659 |
+
"grad_norm": 23.743509110621172,
|
| 660 |
+
"learning_rate": 3.071005917159763e-07,
|
| 661 |
+
"logits/chosen": -1.671875,
|
| 662 |
+
"logits/rejected": -1.65625,
|
| 663 |
+
"logps/chosen": -245.0,
|
| 664 |
+
"logps/rejected": -255.0,
|
| 665 |
+
"loss": 0.3688,
|
| 666 |
+
"rewards/accuracies": 0.862500011920929,
|
| 667 |
+
"rewards/chosen": -0.166015625,
|
| 668 |
+
"rewards/margins": 1.0234375,
|
| 669 |
+
"rewards/rejected": -1.1953125,
|
| 670 |
+
"step": 420
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 1.3738019169329074,
|
| 674 |
+
"grad_norm": 21.0489995042185,
|
| 675 |
+
"learning_rate": 3.011834319526627e-07,
|
| 676 |
+
"logits/chosen": -1.671875,
|
| 677 |
+
"logits/rejected": -1.6953125,
|
| 678 |
+
"logps/chosen": -243.0,
|
| 679 |
+
"logps/rejected": -256.0,
|
| 680 |
+
"loss": 0.3246,
|
| 681 |
+
"rewards/accuracies": 0.875,
|
| 682 |
+
"rewards/chosen": -0.126953125,
|
| 683 |
+
"rewards/margins": 1.2265625,
|
| 684 |
+
"rewards/rejected": -1.359375,
|
| 685 |
+
"step": 430
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.4057507987220448,
|
| 689 |
+
"grad_norm": 23.98752497138385,
|
| 690 |
+
"learning_rate": 2.952662721893491e-07,
|
| 691 |
+
"logits/chosen": -1.6796875,
|
| 692 |
+
"logits/rejected": -1.625,
|
| 693 |
+
"logps/chosen": -242.0,
|
| 694 |
+
"logps/rejected": -264.0,
|
| 695 |
+
"loss": 0.3287,
|
| 696 |
+
"rewards/accuracies": 0.9375,
|
| 697 |
+
"rewards/chosen": -0.26953125,
|
| 698 |
+
"rewards/margins": 1.390625,
|
| 699 |
+
"rewards/rejected": -1.6640625,
|
| 700 |
+
"step": 440
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 1.4376996805111821,
|
| 704 |
+
"grad_norm": 22.151863656806025,
|
| 705 |
+
"learning_rate": 2.893491124260355e-07,
|
| 706 |
+
"logits/chosen": -1.640625,
|
| 707 |
+
"logits/rejected": -1.6953125,
|
| 708 |
+
"logps/chosen": -242.0,
|
| 709 |
+
"logps/rejected": -256.0,
|
| 710 |
+
"loss": 0.3172,
|
| 711 |
+
"rewards/accuracies": 0.887499988079071,
|
| 712 |
+
"rewards/chosen": -0.2470703125,
|
| 713 |
+
"rewards/margins": 1.359375,
|
| 714 |
+
"rewards/rejected": -1.6015625,
|
| 715 |
+
"step": 450
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 1.4696485623003195,
|
| 719 |
+
"grad_norm": 25.183351240662166,
|
| 720 |
+
"learning_rate": 2.834319526627219e-07,
|
| 721 |
+
"logits/chosen": -1.671875,
|
| 722 |
+
"logits/rejected": -1.6953125,
|
| 723 |
+
"logps/chosen": -240.0,
|
| 724 |
+
"logps/rejected": -260.0,
|
| 725 |
+
"loss": 0.3275,
|
| 726 |
+
"rewards/accuracies": 0.925000011920929,
|
| 727 |
+
"rewards/chosen": -0.240234375,
|
| 728 |
+
"rewards/margins": 1.4609375,
|
| 729 |
+
"rewards/rejected": -1.6953125,
|
| 730 |
+
"step": 460
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 1.5015974440894568,
|
| 734 |
+
"grad_norm": 23.9596209798775,
|
| 735 |
+
"learning_rate": 2.7751479289940824e-07,
|
| 736 |
+
"logits/chosen": -1.6953125,
|
| 737 |
+
"logits/rejected": -1.671875,
|
| 738 |
+
"logps/chosen": -240.0,
|
| 739 |
+
"logps/rejected": -258.0,
|
| 740 |
+
"loss": 0.2869,
|
| 741 |
+
"rewards/accuracies": 0.949999988079071,
|
| 742 |
+
"rewards/chosen": -0.232421875,
|
| 743 |
+
"rewards/margins": 1.46875,
|
| 744 |
+
"rewards/rejected": -1.6953125,
|
| 745 |
+
"step": 470
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.5335463258785942,
|
| 749 |
+
"grad_norm": 20.884264225876592,
|
| 750 |
+
"learning_rate": 2.715976331360947e-07,
|
| 751 |
+
"logits/chosen": -1.6875,
|
| 752 |
+
"logits/rejected": -1.65625,
|
| 753 |
+
"logps/chosen": -238.0,
|
| 754 |
+
"logps/rejected": -262.0,
|
| 755 |
+
"loss": 0.3432,
|
| 756 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 757 |
+
"rewards/chosen": -0.1142578125,
|
| 758 |
+
"rewards/margins": 1.5546875,
|
| 759 |
+
"rewards/rejected": -1.671875,
|
| 760 |
+
"step": 480
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 1.5654952076677318,
|
| 764 |
+
"grad_norm": 27.424882575382,
|
| 765 |
+
"learning_rate": 2.6568047337278106e-07,
|
| 766 |
+
"logits/chosen": -1.6015625,
|
| 767 |
+
"logits/rejected": -1.65625,
|
| 768 |
+
"logps/chosen": -250.0,
|
| 769 |
+
"logps/rejected": -264.0,
|
| 770 |
+
"loss": 0.3153,
|
| 771 |
+
"rewards/accuracies": 0.887499988079071,
|
| 772 |
+
"rewards/chosen": -0.5625,
|
| 773 |
+
"rewards/margins": 1.4296875,
|
| 774 |
+
"rewards/rejected": -1.984375,
|
| 775 |
+
"step": 490
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 1.5974440894568689,
|
| 779 |
+
"grad_norm": 25.335220219602046,
|
| 780 |
+
"learning_rate": 2.5976331360946746e-07,
|
| 781 |
+
"logits/chosen": -1.6875,
|
| 782 |
+
"logits/rejected": -1.65625,
|
| 783 |
+
"logps/chosen": -244.0,
|
| 784 |
+
"logps/rejected": -260.0,
|
| 785 |
+
"loss": 0.3238,
|
| 786 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 787 |
+
"rewards/chosen": -0.279296875,
|
| 788 |
+
"rewards/margins": 1.5859375,
|
| 789 |
+
"rewards/rejected": -1.8671875,
|
| 790 |
+
"step": 500
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 1.6293929712460065,
|
| 794 |
+
"grad_norm": 26.281054947397205,
|
| 795 |
+
"learning_rate": 2.538461538461538e-07,
|
| 796 |
+
"logits/chosen": -1.640625,
|
| 797 |
+
"logits/rejected": -1.6875,
|
| 798 |
+
"logps/chosen": -243.0,
|
| 799 |
+
"logps/rejected": -255.0,
|
| 800 |
+
"loss": 0.3293,
|
| 801 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 802 |
+
"rewards/chosen": -0.1962890625,
|
| 803 |
+
"rewards/margins": 1.40625,
|
| 804 |
+
"rewards/rejected": -1.6015625,
|
| 805 |
+
"step": 510
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 1.6613418530351438,
|
| 809 |
+
"grad_norm": 22.709993123443564,
|
| 810 |
+
"learning_rate": 2.479289940828402e-07,
|
| 811 |
+
"logits/chosen": -1.6796875,
|
| 812 |
+
"logits/rejected": -1.671875,
|
| 813 |
+
"logps/chosen": -246.0,
|
| 814 |
+
"logps/rejected": -262.0,
|
| 815 |
+
"loss": 0.306,
|
| 816 |
+
"rewards/accuracies": 0.875,
|
| 817 |
+
"rewards/chosen": -0.1865234375,
|
| 818 |
+
"rewards/margins": 1.4609375,
|
| 819 |
+
"rewards/rejected": -1.6484375,
|
| 820 |
+
"step": 520
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 1.6932907348242812,
|
| 824 |
+
"grad_norm": 24.667533534060208,
|
| 825 |
+
"learning_rate": 2.420118343195266e-07,
|
| 826 |
+
"logits/chosen": -1.6328125,
|
| 827 |
+
"logits/rejected": -1.6875,
|
| 828 |
+
"logps/chosen": -239.0,
|
| 829 |
+
"logps/rejected": -260.0,
|
| 830 |
+
"loss": 0.3221,
|
| 831 |
+
"rewards/accuracies": 0.862500011920929,
|
| 832 |
+
"rewards/chosen": -0.0152587890625,
|
| 833 |
+
"rewards/margins": 1.421875,
|
| 834 |
+
"rewards/rejected": -1.4375,
|
| 835 |
+
"step": 530
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 1.7252396166134185,
|
| 839 |
+
"grad_norm": 16.738356385012416,
|
| 840 |
+
"learning_rate": 2.36094674556213e-07,
|
| 841 |
+
"logits/chosen": -1.6328125,
|
| 842 |
+
"logits/rejected": -1.6875,
|
| 843 |
+
"logps/chosen": -240.0,
|
| 844 |
+
"logps/rejected": -260.0,
|
| 845 |
+
"loss": 0.3271,
|
| 846 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 847 |
+
"rewards/chosen": -0.125,
|
| 848 |
+
"rewards/margins": 1.609375,
|
| 849 |
+
"rewards/rejected": -1.734375,
|
| 850 |
+
"step": 540
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 1.7571884984025559,
|
| 854 |
+
"grad_norm": 19.302611241291174,
|
| 855 |
+
"learning_rate": 2.301775147928994e-07,
|
| 856 |
+
"logits/chosen": -1.6484375,
|
| 857 |
+
"logits/rejected": -1.65625,
|
| 858 |
+
"logps/chosen": -248.0,
|
| 859 |
+
"logps/rejected": -268.0,
|
| 860 |
+
"loss": 0.318,
|
| 861 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 862 |
+
"rewards/chosen": -0.1337890625,
|
| 863 |
+
"rewards/margins": 1.546875,
|
| 864 |
+
"rewards/rejected": -1.6796875,
|
| 865 |
+
"step": 550
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.7891373801916934,
|
| 869 |
+
"grad_norm": 17.493675951115446,
|
| 870 |
+
"learning_rate": 2.242603550295858e-07,
|
| 871 |
+
"logits/chosen": -1.65625,
|
| 872 |
+
"logits/rejected": -1.65625,
|
| 873 |
+
"logps/chosen": -240.0,
|
| 874 |
+
"logps/rejected": -266.0,
|
| 875 |
+
"loss": 0.2807,
|
| 876 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 877 |
+
"rewards/chosen": -0.1884765625,
|
| 878 |
+
"rewards/margins": 1.546875,
|
| 879 |
+
"rewards/rejected": -1.7421875,
|
| 880 |
+
"step": 560
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 1.8210862619808306,
|
| 884 |
+
"grad_norm": 22.291882579304676,
|
| 885 |
+
"learning_rate": 2.1834319526627217e-07,
|
| 886 |
+
"logits/chosen": -1.65625,
|
| 887 |
+
"logits/rejected": -1.671875,
|
| 888 |
+
"logps/chosen": -245.0,
|
| 889 |
+
"logps/rejected": -262.0,
|
| 890 |
+
"loss": 0.2885,
|
| 891 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 892 |
+
"rewards/chosen": -0.1240234375,
|
| 893 |
+
"rewards/margins": 1.7265625,
|
| 894 |
+
"rewards/rejected": -1.8515625,
|
| 895 |
+
"step": 570
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.8530351437699681,
|
| 899 |
+
"grad_norm": 26.161100115274262,
|
| 900 |
+
"learning_rate": 2.1242603550295858e-07,
|
| 901 |
+
"logits/chosen": -1.640625,
|
| 902 |
+
"logits/rejected": -1.6328125,
|
| 903 |
+
"logps/chosen": -246.0,
|
| 904 |
+
"logps/rejected": -268.0,
|
| 905 |
+
"loss": 0.3145,
|
| 906 |
+
"rewards/accuracies": 0.887499988079071,
|
| 907 |
+
"rewards/chosen": -0.287109375,
|
| 908 |
+
"rewards/margins": 1.71875,
|
| 909 |
+
"rewards/rejected": -2.0,
|
| 910 |
+
"step": 580
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 1.8849840255591053,
|
| 914 |
+
"grad_norm": 28.81776072776743,
|
| 915 |
+
"learning_rate": 2.0650887573964496e-07,
|
| 916 |
+
"logits/chosen": -1.6484375,
|
| 917 |
+
"logits/rejected": -1.6484375,
|
| 918 |
+
"logps/chosen": -240.0,
|
| 919 |
+
"logps/rejected": -260.0,
|
| 920 |
+
"loss": 0.3249,
|
| 921 |
+
"rewards/accuracies": 0.862500011920929,
|
| 922 |
+
"rewards/chosen": -0.0234375,
|
| 923 |
+
"rewards/margins": 1.5078125,
|
| 924 |
+
"rewards/rejected": -1.53125,
|
| 925 |
+
"step": 590
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.9169329073482428,
|
| 929 |
+
"grad_norm": 19.977671114909082,
|
| 930 |
+
"learning_rate": 2.0059171597633133e-07,
|
| 931 |
+
"logits/chosen": -1.6171875,
|
| 932 |
+
"logits/rejected": -1.6640625,
|
| 933 |
+
"logps/chosen": -239.0,
|
| 934 |
+
"logps/rejected": -254.0,
|
| 935 |
+
"loss": 0.3448,
|
| 936 |
+
"rewards/accuracies": 0.925000011920929,
|
| 937 |
+
"rewards/chosen": -0.115234375,
|
| 938 |
+
"rewards/margins": 1.421875,
|
| 939 |
+
"rewards/rejected": -1.5390625,
|
| 940 |
+
"step": 600
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 1.9488817891373802,
|
| 944 |
+
"grad_norm": 17.659228575887397,
|
| 945 |
+
"learning_rate": 1.9467455621301774e-07,
|
| 946 |
+
"logits/chosen": -1.671875,
|
| 947 |
+
"logits/rejected": -1.640625,
|
| 948 |
+
"logps/chosen": -247.0,
|
| 949 |
+
"logps/rejected": -266.0,
|
| 950 |
+
"loss": 0.3007,
|
| 951 |
+
"rewards/accuracies": 0.9375,
|
| 952 |
+
"rewards/chosen": -0.162109375,
|
| 953 |
+
"rewards/margins": 1.53125,
|
| 954 |
+
"rewards/rejected": -1.6953125,
|
| 955 |
+
"step": 610
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.9808306709265175,
|
| 959 |
+
"grad_norm": 18.405238339078597,
|
| 960 |
+
"learning_rate": 1.8875739644970412e-07,
|
| 961 |
+
"logits/chosen": -1.65625,
|
| 962 |
+
"logits/rejected": -1.6484375,
|
| 963 |
+
"logps/chosen": -242.0,
|
| 964 |
+
"logps/rejected": -260.0,
|
| 965 |
+
"loss": 0.3216,
|
| 966 |
+
"rewards/accuracies": 0.887499988079071,
|
| 967 |
+
"rewards/chosen": -0.12255859375,
|
| 968 |
+
"rewards/margins": 1.46875,
|
| 969 |
+
"rewards/rejected": -1.5859375,
|
| 970 |
+
"step": 620
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 2.0,
|
| 974 |
+
"eval_logits/chosen": -1.65625,
|
| 975 |
+
"eval_logits/rejected": -1.671875,
|
| 976 |
+
"eval_logps/chosen": -250.0,
|
| 977 |
+
"eval_logps/rejected": -253.0,
|
| 978 |
+
"eval_loss": 0.600781261920929,
|
| 979 |
+
"eval_rewards/accuracies": 0.6428571343421936,
|
| 980 |
+
"eval_rewards/chosen": -0.734375,
|
| 981 |
+
"eval_rewards/margins": 0.44140625,
|
| 982 |
+
"eval_rewards/rejected": -1.171875,
|
| 983 |
+
"eval_runtime": 12.6155,
|
| 984 |
+
"eval_samples_per_second": 15.854,
|
| 985 |
+
"eval_steps_per_second": 0.555,
|
| 986 |
+
"step": 626
|
| 987 |
}
|
| 988 |
],
|
| 989 |
"logging_steps": 10,
|